There has been new progress in research on brain like intelligence. On October 9, the reporter learned from the Institute of Automation of the Chinese Academy of Sciences that inspired by the structure of biological brain neural circuits, the researchers of the Institute proposed a brain inspired autonomous evolution model of artificial intelligence neural circuits. Based on this model, they developed a more biologically reasonable and efficient brain pulse like neural network. The relevant research paper is published online in the Proceedings of the National Academy of Sciences. In the biological nervous system, different types of neurons can self-organize into neural circuits with different connection patterns to support the implementation of rich cognitive functions structurally. The different types of neural circuits and their adaptive abilities in the human brain promote the realization of human perception, learning, decision-making, and other advanced cognitive functions. "However, most of the current design paradigms of impulsive neural networks are based on structures in the field of deep learning. These structures significantly hinder the potential of impulsive neural networks in complex tasks." Zeng Yi, the corresponding author of the paper and a researcher at the Institute of Automation of the Chinese Academy of Sciences, said that if the structure of biological neural loops is applied to the design of brain like impulsive neural networks, the ability of artificial intelligence systems can be greatly improved. Inspired by the diversity of biological brain neural circuit structures that have undergone natural evolution and the mechanism of pulse timing dependent plasticity, Zeng Yi's team utilized the local pulse behavior of neurons to autonomously evolve more biologically rational functional neural circuits through local rules of pulse timing dependent plasticity; At the same time, by incorporating a self generated neural loop, the team constructed a brain pulse like neural network for image classification, reinforcement learning, and decision-making tasks. Zeng Yi stated that the self evolving brain pulse like neural network has achieved performance comparable to artificial neural networks, laying the foundation for the evolution of networks with complex functions and the emergence of cognitive abilities. Our experiments have shown that these structures can better help solve core problems related to artificial intelligence such as learning and decision-making, which provides inspiration for future research on general brain like cognitive intelligence, "explained Zeng Yi. (New News Agency)